Conventionally, there is a digest video image generating apparatus whose object is to provide a new digest video image generation technique for efficiently generating a digest video image containing a scene that is appropriate as a digest scene in a video image and having a length that is within a predetermined period of time (see Patent Document 1).
In this apparatus, one or multiple scene sections that are appropriate as a digest video image and important in terms of a meaning are selected using a given method from an input video image. The value of the time length of the digest video image to be generated is preset and is compared with the total time length of the selected scene sections being important in terms of the meaning. According to the comparison result, scene sections are further added, or on the contrary, partially deleted from previously selected scene sections, thereby defining the digest scene so as to allow the time length of the digest video image to exactly match the preset value, and thus the digest video image is generated.